Python: Longest consecutive-day activity streak per user

Problem

Given event logs where each row is:

user_id,timestamp,event

Compute, for each user, the length of their longest consecutive-day streak with at least one event.

Key idea

For each user:

1. Convert timestamps to calendar dates.
2. Deduplicate dates (set) so multiple events on the same day count once.
3. Sort unique dates.
4. Scan once and count streaks where (d2 - d1).days == 1.

Python snippet

from collections import defaultdict
from datetime import datetime


def longest_streak(dates):
    dates = sorted(set(dates))
    best = cur = 0
    prev = None

    for d in dates:
        cur = cur + 1 if prev and (d - prev).days == 1 else 1
        best = max(best, cur)
        prev = d

    return best


def per_user_streaks(rows):
    # rows: iterable[(user_id, timestamp_str, event)]
    by_user = defaultdict(list)

    for user_id, ts, _event in rows:
        day = datetime.fromisoformat(ts).date()
        by_user[user_id].append(day)

    return {u: longest_streak(days) for u, days in by_user.items()}

Complexity

• Let k = number of unique active days for a user.
• Per user: O(k log k) for sort + O(k) scan.
• Memory: O(k).

Notes

• If timestamps are UTC but streaks should be local-time, convert timezone before taking .date().
• For very large datasets, pre-aggregate by (user_id, date) upstream before computing streaks.